Visualizing rental listings¶
In [1]:
%matplotlib inline
import geopandas as gpd
import pandas as pd
import numpy as np
import pysal as ps
import folium
from shapely.geometry import Point
from folium import IFrame
from folium.plugins import MarkerCluster
Read the rental listings prepared and saved earlier¶
In [2]:
rents_df_new = pd.read_csv('./data/rent/geocoded_rents.csv', sep=',', encoding='utf-8')
In [3]:
rents_df_new.head()
Out[3]:
In [4]:
rents_df_new.loc[pd.isna(rents_df_new['lat1'])].shape
Out[4]:
Convert data into a Geodataframe¶
In [5]:
rents_geom = [Point(xy) for xy in zip(rents_df_new.lon1, rents_df_new.lat1)]
rents_df_temp = rents_df_new.drop(['lat1', 'lon1'], axis = 1)
crs = {'init': 'epsg:4326'}
rents_gdf = gpd.GeoDataFrame(rents_df_temp, crs=crs, geometry=rents_geom)
In [6]:
rents_gdf.plot()
Out[6]:
In [7]:
rents_gdf.shape
Out[7]:
Read in the shape file for Berlin admin regions¶
In [9]:
berlin_file = "./data/Berlin__RBSOrtsteile"
berlin_df = gpd.read_file(berlin_file, encoding='utf-8')
print(type(berlin_df))
berlin_df
Out[9]:
In [10]:
berlin_df.plot()
Out[10]:
In [12]:
berlin_df.shape
Out[12]:
Spatial join to find listings for each region¶
In [11]:
rents_with_districts = gpd.sjoin(berlin_df, rents_gdf)
In [13]:
rents_with_districts.shape
Out[13]:
Compute average rent per square meter for each region¶
In [14]:
avg_rents = gpd.sjoin(berlin_df, rents_gdf).groupby('OBJECTID').agg({'price_sq_m' : np.mean})
In [15]:
rents_with_districts.shape, avg_rents.shape
Out[15]:
In [16]:
avg_rents_with_districts = pd.merge(berlin_df, avg_rents, left_on='OBJECTID', right_on='OBJECTID')
In [17]:
avg_rents_with_districts.head()
Out[17]:
In [18]:
avg_rents_with_districts.plot(column='price_sq_m', cmap='OrRd', scheme='Fisher_Jenks')
Out[18]:
Interactive visualization¶
In [45]:
print('Producing folium choropleth and cluster marker visualization for rents')
# Create a Folium basemap centered in Berlin
rents_map = folium.Map([52.5161,13.3770], zoom_start = 10)
def add_choropleth_to_map(mapobj):
# Generate the list of threshold values for classification based on a Fisher Jenks classifier
my_threshold_scale = ps.esda.mapclassify.Fisher_Jenks(avg_rents_with_districts['price_sq_m']).bins.tolist()
# Call the Folium choropleth function, specifying the geometry from the GeoDataFrame converted to GeoJSON
# and the data as the GeoDataFrame
mapobj.choropleth(geo_data = avg_rents_with_districts.to_json(), data = avg_rents_with_districts, \
columns = ['OBJECTID', 'price_sq_m'], key_on = 'feature.properties.{}'.format('OBJECTID'), \
fill_color = 'YlOrRd', fill_opacity = 0.6, line_opacity = 0.2, \
threshold_scale = my_threshold_scale)
for idx, row in avg_rents_with_districts.iterrows():
popup = row['Ortsteilna'] + "\n" + str(row['price_sq_m'])
#popup = IFrame(popup_label, width = 300, height = 100)
folium.CircleMarker([row.geometry.centroid.y, row.geometry.centroid.x], popup=popup,
radius=2, color='white').add_to(mapobj)
return mapobj
# Add the choropleth to the basemap
rents_map=add_choropleth_to_map(rents_map)
In [46]:
def add_clusters_to_map(mapobj):
# The location coordinates and the pop-ups for the single activity markers
coords, popups = [], []
# The information to be shown in the marker pop-ups
popup_fields = ['url', 'price_sq_m']
for i, row in rents_gdf.iterrows():
# Populate the list of coordinates with the list of activites
coords.append([row.geometry.y, row.geometry.x])
# Create HTML for the IFrame popup
label = '<br>'.join([str(row[field]) for field in popup_fields])
# Append the IFrame containing the HTML to the pop-ups list
popups.append(IFrame(label, width = 300, height = 100))
# Create a Folium feature group, since we will be displaying multiple layers
point_layer = folium.FeatureGroup(name = 'point_layer')
# Add the clustered points and pop-ups to this layer
point_layer.add_child(MarkerCluster(locations = coords, popups = popups))
# Add this point layer to the map object
mapobj.add_child(point_layer)
return mapobj
# Add the clusters layer to the map
rents_map = add_clusters_to_map(rents_map)
In [47]:
folium.LayerControl().add_to(rents_map) # Allows layers to be toggled on/off in the map
rents_map.save('rents_map2.html') # Save map as HTML file
print('Saving visualization under: rents_map.html')
In [48]:
rents_map
Out[48]:
In [ ]: